| Many Web search tools, such as Netscape, AltaVista, Excite, Lycos, and Google, have been developed to provide an efficient and automatic way to retrieve fast-growing, distributed and heterogeneous information. But the user often faces a tedious task to find the most relevant Web sites from the many retrieved hyperlinks generated by these search tools in response to a query. WebRank is a ranked query system which overcomes this by presenting the user with the top-ranked Web pages, and discarding the other ones.;In order to achieve a high quality ranking result, WebRank adopts a set of novel rough set search strategies for ranking the Web pages. The rough set model is a powerful tool for dealing with the vagueness and uncertainty of human knowledge. The rough set Web page ranking approach is based on the notions of rough equality, rough inclusion and rough overlap. It applies a refined, extended rough set model that gives a more precise description of these rough relations for the rough Web attribute sets. (Abstract shortened by UMI.). |
